Plant Foods Hum Nutr (2013) 68:229234

DOI 10.1007/s11130-013-0360-2

ORIGINAL PAPER

Comparative Study of Anthocyanin and Volatile Compounds

Content of Four Varieties of Mexican Roselle (Hibiscus
sabdariffa L.) by Multivariable Analysis
G. A. Camelo-Mndez & J. A. Ragazzo-Snchez &
A. R. Jimnez-Aparicio & P. E. Vanegas-Espinoza &
O. Paredes-Lpez & A. A. Del Villar-Martnez

Published online: 29 May 2013

# Springer Science+Business Media New York 2013

Abstract Anthocyanins are a group of water-soluble pigments

that provide red, purple or blue color to the leaves, flowers, and
fruits. In addition, benefits have been attributed to hypertension
and cardiovascular diseases. This study compared the content
of total anthocyanins and volatile compounds in aqueous and
ethanolic extracts of four varieties of Mexican roselle, with
different levels of pigmentation. The multivariable analysis of
categorical data demonstrated that ethanol was the best solvent
for the extraction of both anthocyanins and volatile compounds.
The concentration of anthocyanin in pigmented varieties ranged
from 17.3 to 32.2 mg of cyanidin 3-glucoside/g dry weight,
while volatile compounds analysis showed that geraniol was
the main compound in extracts from the four varieties. The
principal component analysis (PCA) allowed description of
results with 77.38 % of variance establishing a clear grouping
for each variety in addition to similarities among some of these
varieties. These results were validated by the confusion matrix
obtained in the classification by the factorial discriminate analysis (FDA); it can be useful for roselle varieties classification.
Electronic supplementary material The online version of this article
(doi:10.1007/s11130-013-0360-2) contains supplementary material,
which is available to authorized users.
G. A. Camelo-Mndez : A. R. Jimnez-Aparicio :
P. E. Vanegas-Espinoza : A. A. Del Villar-Martnez (*)
Instituto Politcnico Nacional, CEPROBI, Carretera
Yautepec-Jojutla Km. 6, Calle CEPROBI 8, Col. San Isidro,
Yautepec, Morelos, Mxico C.P. 62731
e-mail: adelvillarm@ipn.mx
J. A. Ragazzo-Snchez
Laboratorio Integral de Investigacin en Alimentos, Instituto
Tecnolgico de Tepic, Tepic, Nayarit, Mxico
O. Paredes-Lpez
Centro de Investigacin y de Estudios Avanzados del IPN,
Unidad Irapuato, Irapuato, Guanajuato, Mxico

Small differences in anthocyanin and volatile compounds content could be detected, and it may be of interest for the food
industry in order to classify a new individual into one of several
groups using different variables at once.
Keywords Roselle . Anthocyanins . Volatile compounds .
Categorical multifactorial analysis . Principal component
analysis (PCA) and factorial discriminate analysis (DFA)

Introduction
Hibiscus sabdariffa L. (Malvaceae) a plant native to Africa is
grown in tropical areas such as Sudan, Taiwan, Thailand and
Mexico [13]. It is a short-day annual shrub that is known by
different synonyms and common names such as roselle in
the U.S. and England, loiselle in France, jamaica or flor
de jamaica in Mexico and Spain, karkade in Sudan and
Arabia [4]. This plant grows over 180 cm; with globular
stems, leaves are ovate, flowers are grouped in terminal clusters, and petals are white with a reddish center at the base of
the staminal column. The calyx grows during maturity, fruit is
fleshy bright red, and the red persistent of calyx is the main
component having a sour taste, commonly used in preparation
of drinks, and as a food coloring [1].
In Mexico, it has recently gained an important position in
the market for soft drinks and commercial preparations.
Currently extracts of H. sabdariffa are marketed as supplements, because of the perception of their potential health
benefits [5, 6]. In traditional medicine it is used to treat kidney
stones [7], as a diuretic and has been studied extensively as an
antispasmodic [8], anti-inflammatory [9], treatment of atherosclerosis [10], antioxidant and anticancer treatment, bactericide and fungicide [5].

230

There are few studies about characterization of roselle

varieties with different degrees of pigmentation. Christian
and Jackson [11] reported that the dark red calyx has
the highest anthocyanin content and antioxidant activity
followed by the red variety, while the green variety does
not present any compound with these properties. It is
reported that the exotic aroma of red roselle infusion is
provided by the interaction of terpenoids with floral notes,
fatty acid with acid notes, and sugar degradation products
with a caramel-like odor [1214]. However, there are no
reported volatile profiles of roselle varieties with different
degrees of pigmentation. Therefore, the aim of this study
was to compare the contents of anthocyanins and volatile
compounds in aqueous and ethanolic extracts of four varieties of roselle (Hibiscus sabdariffa L.) with different degrees
of pigmentation.

Materials and Methods

Plant Material
Roselle calyces were obtained from the experimental
field of the Colegio Superior de Agricultura del Estado
de Guerrero (CSAEGRO) Iguala, Guerrero, Mexico, during 2011 winter season (NovemberDecember). Negra,
Sudan, Rosa and Blanca roselle varieties were used; they
were obtained from breeding plans and named according
to calyx pigmentation by Quintn Obispo Gonzlez.
Dried calyx was lyophilized and pulverized in a food
processor (Sunbeam, 220 watts motor, Kintchen Aid,
Canada), flour was stored in airtight and opaque containers at 4 C.
Total Anthocyanin Extraction
The aqueous extracts were obtained by mixing the flour in
proportion 1:10 [sample:water (w/v)], while the ethanolic
extracts were obtained according to the method reported by
Zhao et al. [15]. 1 g of flour was mixed in 10 ml of acidified
ethanol solution [0.1 M HCl (v/v)]. These mixtures were
stirred for 24 h at 200 rpm under dark conditions at room
temperature, and then filtered through Whatman No. 42
filter paper.
Quantification of Total Anthocyanins Content (TAC)
Anthocyanins are the main pigments in roselle [2, 4, 5], they
were determined by differential pH methods described by
Zhao et al. [15], two dilutions were prepared, one of these in
0.025 M potassium chloride buffer, pH 1.0; and the second
one in sodium acetate buffer 0.4 M, pH 4.5. The absorbance
was measured in spectrophotometer (UV-1601, Shimadzu,

Plant Foods Hum Nutr (2013) 68:229234

Japan) at 510 and 700 nm, respectively. The absorbance was

calculated by Eq. 1:
A A510  A700 pH1:0  A510  A700 pH4:5

the TAC was calculated using Eq. 2, and the results were
expressed as mg of cyanidin 3-glucoside/L.
TAC

AB
D
 MV   V  100
eL
G

Where AB is absorbance; e is the molar extinction coefficient (molar absorptivity) of cyanidin 3-glucoside (26,900)
[18]; L is the cell length (1 cm); MW is anthocyanins
molecular weight (449.2); D is dilution factor; V is final volume
(mL); and G is the dry weight (dw) of roselle flour (mg).
Solid Phase Micro-Extraction (SPME)
For extraction of volatile compounds, ethanolic extracts of
roselle varieties were evaporated to dryness under an inert
nitrogen atmosphere.
A mixture of 1 g of sample and 10 ml of distilled water was
prepared in a 20 ml vial with 5 ml of a saturated NaCl solution
was added [16]. The vial was sealed airtight and incubated at
40 C for 20 min. Carboxen/PDMS of 65 m fiber (Supelco,
Bellefonte, PA) was used for the extraction of volatile compounds, in which the fiber needle was inserted and exposed in
the headspace of the vial for 20 min. Desorption was finally
performed by exposing the fiber for 4 min in the injection port
of the gas chromatograph.
Determination of Volatile Compounds by Gas
Chromatography (GC)
Gas chromatograph Varian 3800 GC (USA) was used and
analysis of volatile compounds was done according to the
methodology proposed by Pino et al. [13]. Separations were
performed using a CP-Wax 52 CB fused silica capillary
column, 30 m length and 0.25 m diameter. The initial
temperature was 60 C (2 min), then subsequently
programmed at 4 C/min to 250 C. Nitrogen was used as
carrier gas at a flow rate of 1 ml/min. Desoprtion temperature and detector was 250 C and injection was performed in
split mode (1:10). For identification of volatile compounds,
ethanol, geraniol, isoamylacetate, linalool, 2-nonanol, benzaldehyde, menthol, ethyl methyl phenylglycidate, and gamma undecalactone were used as reference standards, 2nonanol was used by the internal standard method.
Statistical Analysis
A categorical multifactorial experimental design with two
factors was used: variety (Negra, Sudan, Rosa and Blanca)
and type of extract (aqueous and ethanolic), with the following

Plant Foods Hum Nutr (2013) 68:229234

response variables: anthocyanins, ethanol, geraniol, isoamyl

acetate, linalool, 2-nonanol, benzaldehyde, menthol, ethyl
methyl phenylglycidate, and gamma undecalactone content.
The least significant difference (LSD) with a significance level
of p<0.05 was calculated. Statgraphics Plus version 5.1
(Manugistics, Inc., Rockville, MA, USA) was used.
Principal Component Analysis (PCA)
The principal component analysis (PCA) was performed
using the method proposed by Shyh-Hung et al. [12]. A
data matrix of normalized correlation was used to calculate
the eigenvalues (loadings), eigenvectors and related components with the original variables. In this study, the first three
principal components were plotted to show and highlight
related data in a scatter plot with three axes.
Factorial Discriminate Analysis (FDA)
The goal of the factorial discriminate analysis (FDA) is to
build a matrix that can predict the appropriate information of
quantitative values of the unknown sample [16, 17]. These
values are represented in a correlation matrix that contains
information about the current and expected classifications
performed by a classification system. The correlation matrix
was used to validate the results obtained by the FDA [16,
17]. For this analysis, the Statgraphics Plus version 5.1
(Manugistics, Inc., Rockville, MA, USA) statistical software was used.

Results and Discussion

Total Anthocyanins Determination
Spectrophotometric analysis of extracts showed regions of
maximum absorption in visible region of the spectrum,
where profiles for Negra, Sudan and Rosa varieties were
similar with maximum absorption regions between 520 and
547 nm, these extracts showed maximum absorbance of
0.986, 1.542 and 0.833, respectively. The maximum spectrum was similar according to Giusti and Ping [18] for
delphinidin and the spectrophotometric profile corresponds
to cinnamic acid non-acylated anthocyanins, whereas
Blanca variety showed maximum absorption at 664.5 nm
(0.107 of absorbance) for which anthocyanins were not
detected; since chlorophyll a is detected at this wavelength
[19, 20], it suggested that this compound might be the
pigment present in this variety.
Anthocyanin concentration in roselle extracts was determined spectrophotometrically using the equation mentioned above. Ethanolic extract of Sudan variety had the
highest anthocyanin content (32.19 0.27 mg cyanidin

231

3-glucoside/100 g dw) compared to the other varieties

with both solvents, this correlates to maximum absorbance values above mentioned, due to the high roselle
calix pigmentation. Aqueous extract of Rosa variety
showed the highest anthocyanin content (17.61.71 mg
cyanidin 3-glucoside/100 g dw).
Statistical results of the categorical multivariate analysis
indicated significant differences depending on the extraction
solvent used (p<0.05), ethanol extracted more efficiently
these compounds in comparison with the aqueous extract.
Furthermore, based on variety, there are significant differences (p<0.05) between Sudan and Blanca variety, with
Sudan variety presenting the highest anthocyanin content
and Blanca variety the lowest. However, there was no significant difference (p<0.05) between Negra and Rosa varieties, suggesting that anthocyanin content is similar between
them. Moreover, interaction between the two factors (solvent and variety) indicated that the ethanol allowed higher
anthocyanins extraction for Sudan and Negra varieties because of their polar character since most of anthocyanins
occur naturally as glucosides, and those structures are more
polar than their corresponding aglycones, nevertheless scientific studies reported that greater efficiency is found in the
hydro-alcoholic extraction at low pH [18].
These results were similar to those reported by Christian
and Jackson [11] in varieties from Jamaica, where they
found that highly pigmented varieties had higher monomeric anthocyanin content, and the unpigmented variety lacked
them. This difference can be attributed to the varieties since
anthocyanin content depends on factors as: varietal genetic
pool and environmental condition (temperature, light, soil
nutrient deficiencies and pathogens attack) [2123].
Analyzed varieties in this study were obtained recently by
traditional breeding in order to improve pigment content, so
this work represents the first biochemical characterization of
these materials.
Volatile Compounds Determination
Extraction of volatile compounds was carried out using the
solid phase micro-extraction technique reported as selective
and efficient in the extraction of volatile and aromatics compounds [17]. To compare the headspace volatiles in the extraction, peak areas were normalized (100 %) to the sample
with the major relative total area. For the aqueous extraction,
Rosa variety presented the highest peak area, followed by
Blanca (13.02 %), Sudan (11.36 %) and Negra (7.68 %). On
the other hand, considering Blanca as 100 %, Rosa represented 74.98 %, Negra 60.83 % and Sudan 59.46 % for ethanolic
extractions. This analysis allowed a comparison of total areas
of the obtained extracts, suggesting that ethanol extracted
more efficiently the volatile compounds than water, this was
confirmed with the multivariate categorical analysis.

232

Nine volatile compounds were identified and quantified

using 2-nonanol as internal standard (y=15697x6688.3;
R2 =0.9969) (Online Resource 1). Ethanol extracted more
compounds in Rosa variety compared to the other samples.
It was determined that geraniol was the most abundant compound in the two types of extracts in comparison with the
other identified compounds. Aqueous extract from Rosa variety presented also high content of ethanol, geraniol, menthol,
2-nonanol, benzaldehyde, linalool, gamma undecalactone,
and ethyl methyl phenylglycidate, compared with the other
varieties. Statistical analysis showed significant differences
(p<0.05) by type of variety and type of solvent used for the
compounds identified, but the interaction of the factors did not
show significant differences in the range undecalactone
(Online Resource 2).
Geraniol concentrations (7.109.73 g/g) in aqueous extraction were higher than those reported for roselle beverages
by Shyh-Hung et al. [12] and Ramrez-Rodrigues et al. [14],
but lower than Pino et al. [13]. Differences on identified
compounds and concentrations might be attributed to solvent
polarity, extraction conditions (time, solute concentration, and
temperature), and process analysis [24, 25]. Terpenoids concentration (geraniol and linalool) was higher than those
reported by Du et al. [26], who analyzed volatile composition
and odor-activity value of Black Diamond and Marion blackberries. Those compounds such as linalool and geraniol, could
be pointed as a part of floral flavor notes, as previously
reported for roselle varieties [12, 13, 27].
PCA and FDA of Volatile Compounds
Concentrations of volatile compounds were different in all
varieties depending on type of extraction. Because the analysis of variance did not allow volatile characterization of the
varieties, it was necessary to use the PCA and FDA. In this
study, both PCA and FDA were used to determine differences
between varieties of Hibiscus and type of extract used.
Recently, PCA has been used to determine some flavors in
roselle varieties [13] and other products [16, 17, 25];
Fig. 1 Principal component
analysis of Roselle extracts.
PC1: principal component 1;
PC2: principal component 2;
PC3: principal component 3.
Aqueous extracts: AB: Blanca
variety, AN: Negra variety, AR:
Rosa variety, AS: Sudan
variety. Ethanolic extracts: EB:
Blanca variety, EN: Negra
variety, ER: Rosa variety, ES:
Sudan variety

Plant Foods Hum Nutr (2013) 68:229234

this technique reduces the dimensionality of a data set.

Intuitively, this technique is used to find the causes of variability in a data set and sort them by importance. Because
PCA is a descriptive method, we used FDA to take a decision
based on quantitative data obtained from the response variables. In this study, the first three principal components (PC1,
PC2, PC3) were analyzed having eigenvalues greater than 1.0
[12]; explaining 43.70, 21.73 and 11.95 % of the variance,
respectively. Hence, together they explain about 77.34 % of
the total variation model. It was found that isoamyl acetate
was the most related variable for PC1; isoamyl acetate, menthol, linanool, ethyl methyl phenylglycidate for PC2; and
isoamyl acetate, menthol, 2-nonanol and benzaldehyde for
PC3. Differences among groupings for varieties based on
extract type could be observed, the PCA analysis showed that
Negra and Sudan varieties (aqueous extract) overlapped with
each other meaning that these two samples showed similarities
among the results obtained for anthocyanins content and
volatile compounds (Fig. 1). In the PCA, about 77 % of the
information was considered in the three main components as
their differences are clearly marked when using a third dimension or third component. PCA established a clear grouping of
each variety and characterization of the tested varieties, as
shown in Fig. 1. Discriminant functions were used to classify
derived observations, and the three highest scores allowed
fitting the model, as well as for any new observation. Scatter
plot showed a quite good varietal separation among samples.
Each group represented an acceptable quality, intensity or
source descriptor. Each matrix group was based on the new
variables. Unknown samples were projected onto the discriminative subspace and compared with each group through the
associated matrix. This procedure is designed to develop a set
of discriminant functions that can help predict extract values
based on other quantitative variables. Twenty four cases were
used to develop a model to differentiate between the eight
levels of extract. Six discriminant functions with p-values less
than 0.05 are statistically significant, additional data are given
in Online Resource 3, 4. This method has been used to predict
quantitative values based on a calibration curve, and is

Plant Foods Hum Nutr (2013) 68:229234

correlated with characteristics of the quantitative variables of

the associated sample [16]. Relative magnitude of the coefficients (Online Resource 4) determine their use as independent
variables to differentiate between groups. This can be done by
cross-validation of the FDA established confusion matrix
(data not showed). Observations were correctly classified
(100 %) for each variety and solvent, obtained from the
confusion matrix, although Fig. 1 showed that some of the
samples overlapped, the FDA confirmed that anthocyanins
and volatiles were characteristics of each variety. There is no
doubt that these results are produced by the selectivity of the
solid phase micro-extraction (SPME), as it eliminates many of
the variables that interfere with organic solvents during extraction [16, 17]. In addition, this technique can detect small
differences (nanograms) in the volatile content between
varieties [16, 17, 25].

Conclusions
This study demonstrated evidence of differences between
anthocyanin and volatile compound contents of four varieties
with different degrees of pigmentation. Ethanol extracted
higher anthocyanins and volatile compounds, showing that
the major volatile compound found in both extracts was
geraniol. PCA and FDA allowed description of about 77 %
of variance values, allowing a clear classification of roselle
varieties using the first three components in a third dimension
scater plott. This model was supported by FDA correlation
confusion matrix. This could be useful for varieties classification according to their anthocyanin and volatile content and,
application of these techniques for natural and additive free
food industry is recommended.
Acknowledgments The authors thank to Colegio Superior de
Agricultura del Estado de Guerrero (CSAEGRO) and Quintn Obispo
Gonzlez for biological samples. One of the authors (GACM) also
acknowledges the scholarship from CONACYT-Mexico and SIP-IPN.
Conflict of interest The authors declare that they have no conflict of
interest.

References
1. Chang YC, Huang HP, Hsu JD, Yang SF, Wang CJ (2005) Hibiscus anthocyanins rich extract-induced apoptotic cell death in human promyelocytic leukemia cells. Toxicol Appl Pharmacol
205:201212
2. Hirunpanich V, Utaipat A, Morales NP, Bunyapraphatsara N, Sato H,
Herunsalee A, Suthisisang C (2005) Antioxidant effects of aqueous
extracts from dried calyx of Hibiscus sabdariffa L. (Roselle) in vitro
using rat low-density lipoprotein (LDL). Biol Pharm Bull 28:
477481
3. Lin HH, Chen JH, Kuo WH, Wang C (2007) Chemopreventive
properties of Hibiscus sabdariffa L. on human gastric carcinoma

233
cell through apoptosis induction and JNK/p38 MAPK signaling
activation. Chem Biol Interact 165:5975
4. Gomes Maganha E, da Costa Halmenschlager R, Moreira Rosa R,
Pegas Henriques JA, Lia de Paula Ramos AL, Saffi J (2010)
Pharmacological evidences for the extracts and secondary metabolites from plants of the genus Hibiscus. Food Chem 118:110
5. Ali BH, Al Wabel N, Blunden G (2005) Phytochemical, pharmacological and toxicological aspects of Hibiscus sabdariffa L.: a
review. Phytother Res 19(5):369375
6. Carvajal O, Waliszewski SM, Barradas DM, Orta Z, Hayward PM,
Nolasco C, Angulo O, Sanchez R, Infanzon RM, Trujillo PRL
(2005) The consumption of Hibiscus sabdariffa dried calyx
ethanolic extract reduced lipid profile in rats. Plant Foods Hum
Nutr 60:153159
7. Yao LH, Jiang YM, Shi J, Toms-Barbern FA, Datta N,
Singanusong R, Chen SS (2004) Flavonoids in food and their
health benefits. Plant Foods Hum Nutr 59(3):113122
8. Ali MB, Salih WM, Mohamed AH, Homeida AM (1991) Investigation of the antispasmodic potential of Hibiscus sabdariffa calyces. J Ethnopharmacol 31:249257
9. Dafallah AA, Al-Mustafa Z (1996) Investigation of the
antiinflamatory activity of Acacia nilotica and Hibiscus sabdariffa.
Am J Chin Med 24:263269
10. Chen CC, Hsu JD, Wang SF, Chiang HC, Yang MY, Kao ES, Ho
YC, Wang CJ (2003) Hibiscus sabdariffa extract inhibits the
development of atherosclerosis in cholesterol-fed rabbits. J Agric
Food Chem 51:54655472
11. Christian KR, Jackson JC (2009) Changes in total phenolic and
monomeric anthocyanin composition and antioxidant activity of
three varieties of sorrel (Hibiscus sabdariffa) during maturity. J
Food Compost Anal 22:663667
12. Shyh-Hung C, Tzou-Chi H, Chi-Tang H, Pi-Jen T (1998) Extraction, analysis, and study on the volatiles in roselle tea. J Agric
Food Chem 46:11011105
13. Pino JA, Mrquez E, Marbot R (2006) Volatile constituents from
tea of roselle (Hibiscus sabdariffa L.) Revista CENIC. Ciencias
Qumicas 37:127129, ISSN-2221-2450
14. Ramrez-Rodrigues MM, Balaban MO, Marshall MR, Rouseff RL
(2011) Hot and cold water infusion aroma profiles of Hibiscus
sabdariffa: fresh compared with dried. J Food Sci 76:212217
15. Zhao X, Corrales M, Zhang C, Hu X, Ma Y, Tauscher B (2008)
Composition and thermal stability of anthocyanins from chinese purple corn (Zea mays L.). J Agric Food Chem 56:10761
10766
16. Solis-Solis HM, Calderon-Santoyo M, Gutierrez-Martinez P,
Schorr-Galindo S, Ragazzo-Sanchez JA (2007) Discrimination of
eight varieties of apricot (Prunus armeniaca) by electronic nose,
LLE and SPME using GC-MS and multivariate analysis. Sens
Actuator B Chem 125:415421
17. Sols-Sols HM, Caldern-Santoyo M, Schorr-Galindo S, LunaSolano G, Ragazzo-Snchez JA (2007) Characterization of aroma
potential of apricot varieties using different extraction techniques.
Food Chem 105:829837
18. Giusti MM, Ping P (2008) Analysis of anthocyanins. In: Cocaciu C
(ed) Food colorants: chemical and functional properties. Taylor &
Francis, Cliy-Napoca
19. Pascal A, Gradinaru C, Wacker U, Peterman E, Calkoen F, Irrgan K,
Horton P, Ranger G, van Grondelle R, Robert B, van Amerongen H
(1999) Spectroscopic characterization of the spinach Lhb4 protein
(CP29), a minor light-harvesting complex of photosystem II. Eur J
Biochem 262:817823
20. Butnariu M (2012) An analysis of Sorghum halepenses behavior
in presence of tropane alkaloids from Datura stramonium extracts.
Chem Cent J 6:75
21. Torres-Morn MI, Escoto-Delgadillo M, Ron-Parra J, Parra-Tovar
G, Mena-Mungua S, Rodrguez-Garca A, Castellanos-Hernndez

234
YO (2011) Relationships among twelve genotypes of roselle (Hibiscus sabdariffa L.) cultivated in western Mexico. Ind Crops Prod
34:10791083
22. Cheng Z, Lu BR, Baldwin BS, Sameshima K, Chen JK (2002)
Comparative studies of genetic diversity in kenaf (Hibiscus
cannabinus L.) varieties based on analysis of agronomic and
RAPD data. Hereditas 136:231239
23. Chen S, Huang T, Ho C, Tsai P (1998) Extraction, analysis, and
study on the volatiles in roselle tea. J Agric Food Chem 46(3):1101
1105
24. Hidalgo-Villatoro SG, Cifuentes-Reyes WAL, Ruano-Sols HH,
Cano-Castillo LE (2009) Caracterizacin de trece genotipos de rosa

Plant Foods Hum Nutr (2013) 68:229234

de jamaica Hibiscus sabdariffa en Guatemala. Agron
Mesoamericana 20:101109, ISSN-1659-1321
25. Gonzalez-Palomares S, Estarrn-Espinosa M, Gmez-Leyva JF,
Andrade-Gonzlez I (2009) Effect of the temperature on the spray
drying of roselle extracts (Hibiscus sabdariffa L.). Plant Foods
Hum Nutr 64:6267
26. Du X, Finn CE, Qian MC (2010) Volatile composition and odouractivity value of thornless Black diamond and Marion blackberries. Food Chem 119(3):11271134
27. Ragazzo-Sanchez JA, Chalier P, Chevalier D, Calderon-Santoyo M,
Ghommidh C (2008) Identification of different alcoholic beverages
by electronic nose coupled to GC. Sens Actuator B Chem 134:4348